On 22.05.2015 (19:44), Gregor Cramer wrote:
> What is the optimal search order? 

With optimal, do you mean computationally or for the user? I.e. how most
efficiently to find the right subset of positions from a huge database, or
how to find the one line or set of lines the user is most likely looking for?

> This means that method (a) has the following behavior:
> 
> 1. Search position in main line
> 2. If (1) fails: Search position in first variation of main line
> 3. If (2) fails: Search position in second variation of main line
> and so on
> 
> And method (b) works in this way:
> 
> 1. Search position in main line (depth 0)
> 2. If (1) fails: Search position in all variations at depth 1
> 3. If (2) fails: Search position in all variations at depth 2
> and so on

My hunch is that the depth-based search would be closer to what the user
would need: a changed move-order is a minor deviation, whereas the order
or variations is - possibly - a fairly random criterion.

> Another method is also possible:
>
> 1. Search in current line until a variation is starting
> 2. Search in variation as soon as a variation is starting
> 3. After end of variation continue search after this variation
> 
> But in this case it may happen that a position will be found in a sub-
> variation, although the same position exists inside the main line. This
> seems to be unexpected.

Do you mean that, owing to different move orders, the sought-for position
will be found first in a variation and only then in the main line? If so,
I agree.


Eyolf

-- 
my eternal youth ornament
just three cents
of emporer's pine

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Scidb-users mailing list
Scidb-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scidb-users

Reply via email to